Self Training Autonomous Driving Agent
This addresses the problem of inefficient training for autonomous driving agents, though it appears incremental as it builds on the World Model architecture.
The paper tackles autonomous driving by proposing a reinforcement learning agent that learns without human assistance, achieving the same accuracy as the original World Model while requiring 96% fewer total agents, 87.5% fewer agents per generation, 70% fewer generations, and 90% fewer rollouts.
Intrinsically, driving is a Markov Decision Process which suits well the reinforcement learning paradigm. In this paper, we propose a novel agent which learns to drive a vehicle without any human assistance. We use the concept of reinforcement learning and evolutionary strategies to train our agent in a 2D simulation environment. Our model's architecture goes beyond the World Model's by introducing difference images in the auto encoder. This novel involvement of difference images in the auto-encoder gives better representation of the latent space with respect to the motion of vehicle and helps an autonomous agent to learn more efficiently how to drive a vehicle. Results show that our method requires fewer (96% less) total agents, (87.5% less) agents per generations, (70% less) generations and (90% less) rollouts than the original architecture while achieving the same accuracy of the original.